Providing intelligence to evolutionary computational methods

Abstract

It is presented an intelligent evolutionary method to solve single objective optimization problems, we called this method Single Objective Intelligent Evolutionary Algorithm (SO-IEA). This method uses several mechanisms that work synergistically to provide the optimal solution by handling in an intelligent way, the number of times that the objective function needs to be evaluated. The SO-IEA was subjected to several tests using complex benchmark functions and the results were statistically compared to other state of the art evolutionary algorithms (EA) obtaining that the SO-IEA outperformed in time and in precision the other methods. In general, the ideas presented here can be easily adapted to other EAs. © 2007 Springer-Verlag Berlin Heidelberg.